meshoptimizer VS libzim

1) I used meshoptimizer to simplify meshes and building meshlets. But to combine these meshlets (for more efficient simplifying) I built a graph with weights and partition it. In this weight some metrics are included: the shared border length of a pair of meshlets, their facing direction and other similar metrics. The next step is to simplify these groups of meshlets (fixate borders for each of them and simplify meshlets within the group). So, you can make dependencies of newly generated meshlets on the initial ones.

Here are some amazing tools for compressing GLTFs: https://github.com/zeux/meshoptimizer/blob/master/gltf/README.md

I've also had good results with this tool: https://github.com/zeux/meshoptimizer/blob/master/gltf/README.md and kt2tdecoder + meshopt_decoder

Implemented Mesh LOD system. LOD selection is done completely on GPU. I generate LODs using meshoptimizer.

Maybe Meshoptimizer can be useful?

Added NVidia mesh shader pipeline support, with old pipeline still supported on top. Meshlets are generated using meshoptimizer.

I've used the "Simplification" operation of meshoptimizer to simplify 3D model geometry: https://github.com/zeux/meshoptimizer

I call meshopt_simplify(). It uses a quadric error metric based mesh simplification of some sort. I haven't really looked into the details of how it works, but the code (with some comments referencing papers) can be found here: https://github.com/zeux/meshoptimizer/blob/master/src/simplifier.cpp

hi! i am building a game engine with c and raylib and i want to use this library for optimizing meshes : https://github.com/zeux/meshoptimizer

Have you considered this? https://github.com/zeux/meshoptimizer

I meant the Kiwix dump (https://download.kiwix.org/zim/wikipedia_en_all_nopic.zim – careful, 60GB!).

At a first glance, the Wikimedia XML dump does not look substantially different from what Kiwix/ZIM does with compressed HTML: They're both compressed (bz2 for the Wikimedia dump, zstd or LZMA for Kiwix/ZIM), and both compress multiple files at once, so inter-file redundancy should hopefully be significantly reduced.

HTML seems a bit more verbose than the Mediawiki syntax (plus the XML header for each article), but I'd be surprised if that actually accounted for a 3x difference in size.

Then again, Kiwix seems to have experimented with shared dictionary brotli compression, which supposedly yields an >2x improvement: https://github.com/openzim/libzim/issues/144

I wonder if their current zstd implementation also uses shared dictionaries. If not, that might just be the reason: If ZIM compression chunks are much smaller than the bz2 streams of the Wikimedia dumps, there would still be a lot of redundancy between chunks.

This is low-level stuff, but the new release of the backend that powers Kiwix clients, libzim 8.1.0 (changelog), now targets WebAssembly (a.k.a. WASM). This is big because it potentially brings the full powers of the Kiwix backend to Web apps such as Kiwix JS and Kiwix JS PWA/Electron. Up till now, these apps have relied on a custom JS port of parts of libzim, but it lacked full features, such as Full Text search. So, you will begin to see us integrating new features into the Web versions soon!

I'm doing my part o7

It's seriously a very interesting and useful dataset that you can do a lot of fun stuff with, if you grab one of the zims without pictures it's of very manageable size too of just a few dozen gigabytes compressed, and there are reasonably good library support in many languages.

Last point doesn't go for Java. Only one I could find for that was this <https://github.com/openzim/libzim>, it's antique and extremely poorly optimized and lacks support for newer compression schemes. I have fixed the performance and added support for zstd compression, but not published the code as it's extremely not finished and major features in the original codebase are very broken. I'll get around to sharing the code some day but right now it's basically permanently mid surgery as I've only patched so far as to get it to extract all or specific files. If anyone wants a copy of this code regardless of state, give me a holler.